Remove buffer pool for tiled GEMM (JuliaLang#42309)

stdlib/LinearAlgebra/src/LinearAlgebra.jl

@@ -586,9 +586,6 @@ function __init__()
             BLAS.lbt_forward(liblapack_path)
         end
         BLAS.check()
-        Threads.resize_nthreads!(Abuf)
-        Threads.resize_nthreads!(Bbuf)
-        Threads.resize_nthreads!(Cbuf)
     catch ex
         Base.showerror_nostdio(ex, "WARNING: Error during initialization of module LinearAlgebra")
     end

stdlib/LinearAlgebra/src/matmul.jl

@@ -726,10 +726,6 @@ function generic_matmatmul(tA, tB, A::AbstractVecOrMat{T}, B::AbstractMatrix{S})
 end
 
 const tilebufsize = 10800  # Approximately 32k/3
-# per-thread arrays of buffers resized by __init__ if needed
-const Abuf = [Vector{UInt8}(undef, tilebufsize)]
-const Bbuf = [Vector{UInt8}(undef, tilebufsize)]
-const Cbuf = [Vector{UInt8}(undef, tilebufsize)]
 
 function generic_matmatmul!(C::AbstractMatrix, tA, tB, A::AbstractMatrix, B::AbstractMatrix,
                             _add::MulAddMul=MulAddMul())
@@ -775,9 +771,8 @@ function _generic_matmatmul!(C::AbstractVecOrMat{R}, tA, tB, A::AbstractVecOrMat
     @inbounds begin
     if tile_size > 0
         sz = (tile_size, tile_size)
-        # FIXME: This code is completely invalid!!!
-        Atile = unsafe_wrap(Array, convert(Ptr{T}, pointer(Abuf[Threads.threadid()])), sz)
-        Btile = unsafe_wrap(Array, convert(Ptr{S}, pointer(Bbuf[Threads.threadid()])), sz)
+        Atile = Array{T}(undef, sz)
+        Btile = Array{S}(undef, sz)
 
         z1 = zero(A[1, 1]*B[1, 1] + A[1, 1]*B[1, 1])
         z = convert(promote_type(typeof(z1), R), z1)
@@ -797,8 +792,7 @@ function _generic_matmatmul!(C::AbstractVecOrMat{R}, tA, tB, A::AbstractVecOrMat
                 end
             end
         else
-            # FIXME: This code is completely invalid!!!
-            Ctile = unsafe_wrap(Array, convert(Ptr{R}, pointer(Cbuf[Threads.threadid()])), sz)
+            Ctile = Array{R}(undef, sz)
             for jb = 1:tile_size:nB
                 jlim = min(jb+tile_size-1,nB)
                 jlen = jlim-jb+1